Crypto small-alt tail (Tidewater)

In plain terms

On the long tail of small altcoins, a market-neutral machine-learning model ranks names and goes long the strongest against the weakest — designed so its small capacity is the moat, not a flaw.

How it works

Technical, funding, on-chain, and sentiment features are fused into a cross-sectional ranking, traded long / short via decentralised perpetual futures, with the whole design pre-registered before testing.

What it’s tested against

A pre-registered three-stage design — train, out-of-sample, then a locked blind holdout run by an independent check — with continuous BTC/ETH beta decomposition and a placebo, so the verdict cannot be massaged.

Data

Crypto market, funding, on-chain, and sentiment data.

Researched — a pre-registered, blind-holdout design for the small-alt tail.